Strain mapping for performance and failure prediction in composites using digital image correlation

Application of loads at the structural level generates spatially varying strain profiles at the micro and meso length scales in advanced composites. These complex deformation patterns influence the progression of damage mechanisms under both static and dynamic conditions. While many studies have attempted to account for architectural effects in material response the precise nature of the conditions that lead to matrix cracking remain somewhat obscure, primarily because of the inability to measure the causative strain distributions. Resultant 'design allowable' values are then derived from exhaustive testing with material performance variation being accounted for through conservative statistical protocols. This paper presents a new application of Digital Image Correlation to make real time, in-situ, spatially varying strain measurements. These measurements show the evolution of deformation in composite architectures, up to the point of matrix cracking and capture the redistribution of strain following these initial failures. The paper also presents work on the representation of experimental observations in an FEA model and shows how Multi-Variate Discriminant Analysis can be used to efficiently identify critical microstructural features that influence performance. In conclusion, it is argued that such techniques have the potential to radically alter, and greatly enhance the efficiency of, material qualification methodologies.